<html><head><style type='text/css'>p { margin: 0; }</style></head><body><div style='font-family: arial,helvetica,sans-serif; font-size: 10pt; color: #000000'>Could be great o have on the Engine:<br>- An upload option for the ISO files<br>- A backup and restore option<br>- An high availability for the engine: install the engine on 2 platforms (hardware?), than integrate them for synchronization<br><br>Jose<br><br><hr id="zwchr"><div style="color: rgb(0, 0, 0); font-weight: normal; font-style: normal; text-decoration: none; font-family: Helvetica,Arial,sans-serif; font-size: 12pt;"><b>From: </b>"noc" <noc@nieuwland.nl><br><b>Cc: </b>users@ovirt.org<br><b>Sent: </b>Sexta-feira, 6 de Setembro de 2013 10:28:09<br><b>Subject: </b>Re: [Users] so, what do you want next in oVirt?<br><br>On 6-9-2013 10:12, Itamar Heim wrote:<br>> On 09/05/2013 10:30 AM, noc wrote:<br>>>>> On 08/21/2013 12:11 PM, Itamar Heim wrote:<br>>>>>> On 08/21/2013 02:40 AM, Joop van de Wege wrote:<br>>>>>>><br>>>>>>> What I would like to see in the ! next version is pxe boot of the<br>>>>>>> nodes.<br>>>>>>> Probably not easy to achieve because of dependency on dhcp.<br>>>>>><br>>>>>> Hi Joop,<br>>>>>><br>>>>>> can you please give a bit more information on the use case / how you<br>>>>>> envision this?<br>>>>>><br>>>>>> current thinking around bare metal provisioning of hosts is to extend<br>>>>>> the functionality around the foreman provider for this, but you may<br>>>>>> have other suggestions?<br>>>>><br>>>>> I think Joop means to be able to add hosts (nodes) to a cluster by<br>>>>> adding their MAC address to the dhcp list for PXE boot into ovirt-node<br>>>>> and thus join the cluster. This would make it easy to add new physical<br>>>>> nodes without any spinning disks or other local storage requirements.<br>>>><br>>>> we started adding foreman integration in 3.3:<br>>>> http://www.ovirt.org/Features/ForemanIntegration<br>>>><br>>>> adding ohad and oved for their thoughts on this.<br>>>><br>>>>><br>>>>> I suppose this may not be easy with complex network connections (bonds<br>>>>> on mgmt network, mgmt network on a tagged vlan, etc), but it should be<br>>>>> possible if the management network interface is plain and physical.<br>>>>><br>>>>> /Simon<br>>>>><br>>>>> PS, Perhaps Joop can confirm this idea, we've talked about it IRL.<br>>>>> _______________________________________________<br>>>>> Users mailing list<br>>>>> Users@ovirt.org<br>>>>> http://lists.ovirt.org/mailman/listinfo/users<br>>>><br>>> This isn't about provisioning with Foreman. Its about having the compute<br>>> nodes NOT having any spinning disks. So the only way to start a node is<br>>> to pxeboot it and then let it (re)connect with the engine. Then it will<br>>> be identified by engine as either a new node or a reconnecting node and<br>>> it will get its configuration from the engine. For reference: thats how<br>>> VirtualIron works. It has a managment network, just like ovirt, and on<br>>> that it runs a tftp and dhcp server. Nodes are plugged into the<br>>> managment network, without disk, and then pxe booted after which they<br>>> appear in the webui as new unconfigured nodes. You then can set various<br>>> settings and upon rebooting the nodes will recieve these settings<br>>> because it is recognised by its mac address. The advantage of this<br>>> construct is that you can place a new server into a rack, cable it,<br>>> power on and go back to you office where you'll find the new node<br>>> waiting to be configured. No messing around with CDs to install an OS,<br>>> not being in the datacenter for hours on end, just in and out.<br>>><br>>> Yes, disks are cheap but they brake down, need maintenance, means<br>>> downtime and in general more admin time then when you don't have them. (<br>>> its a shame to have a raid1 of 2 1Tb disk just to install an OS of less<br>>> then 10G)<br>><br>> just wondering, how do they prevent a rogue node/guest from <br>> masquerading as such a host, getting access/secrets/VMs to be launched <br>> on such an untrusted node (they could easily report a different mac <br>> address if the layer 2 isn't hardened against that)?<br>><br>They would need physical access to your rack which ofcourse is locked, <br>you would need to powerdown/up which would trigger an alert, switch port <br>down/up would trigger an alert, so probably you're notified that <br>something not quite right is happening. I haven't gone through the <br>source to see if there is more then just the mac address check.<br><br>> other than that, yes. we actually used to have this via the <br>> AutoApprovePatterns config option, which would have the engine approve <br>> a pending node as it registers (I admit i don't think anyone used this <br>> last several years, and it may be totally broken by now).<br>><br>> please note this doesn't solve the need for a disk, just the <br>> auto-registration part (if it still works)<br>What I would like is to have the ovirt Node pxe booting and getting its <br>config from engine or autoregister. I know there is a script which <br>converts the iso into a huge pxeboot kernel but don't know how to solve <br>or if its solved the config part.<br><br>@karli:<br>If you run your cluster in Memory Optimization=None then you won't need <br>swap. Have been doing that for years and haven't had a single problem <br>attributed to that. I just would like to have the choice, pxe boot the <br>node and know that you don't have swap. Run with disks if you really <br>need overprovisioning.<br><br>Regards,<br><br>Joop<br><br>_______________________________________________<br>Users mailing list<br>Users@ovirt.org<br>http://lists.ovirt.org/mailman/listinfo/users<br></div><br></div></body></html>